Arctigenin Suppresses Renal Interstitial Fibrosis in a Rat Model of Obstructive

Phytomedicine 30 (2017) 28–41

Contents lists available at ScienceDirect

Phytomedicine

journal homepage: www.elsevier.com/locate/phymed

Arctigenin suppresses renal interstitial ﬁbrosis in a rat model of obstructive nephropathy

Ao Li a,b,1, Xiaoxun Zhang a,1, Mao Shu a, Mingjun Wu c, Jun Wang a, Jingyao Zhang a, ∗ ∗ Rui Wang a, , Peng Li b, , Yitao Wang b a College of Pharmacy and Bioengineering, Chongqing University of Technology, Chongqing 40 0 054, China b State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Macau 999078, China c Institute of Life Sciences, Chongqing Medical University, Chongqing 40 0 016, China

a r t i c l e i n f o a b s t r a c t

Article history: Background: Renal tubulointerstitial ﬁbrosis (TIF) is commonly the ﬁnal result of a variety of progressive Received 28 March 2016 injuries and leads to end-stage renal disease. There are few therapeutic agents currently available for

Revised 28 December 2016 retarding the development of renal TIF. Accepted 9 March 2017 Purpose: The aim of the present study is to evaluate the role of arctigenin (ATG), a lignan component derived from dried burdock ( Arctium lappa L.) fruits, in protecting the kidney against injury by unilateral Keywords: ureteral obstruction (UUO) in rats. Arctigenin Methods: Rats were subjected to UUO and then administered with vehicle, ATG (1 and 3 mg/kg/d), or

Renal ﬁbrosis losartan (20 mg/kg/d) for 11 consecutive days. The renoprotective effects of ATG were evaluated by histo-

Inﬂammation logical examination and multiple biochemical assays. Transforming growth factor- β1 Results: Our results suggest that ATG signiﬁcantly protected the kidney from injury by reducing tubular Oxidative stress

Epithelial-mesenchymal transition dilatation, epithelial atrophy, collagen deposition, and tubulointerstitial compartment expansion. ATG administration dramatically decreased macrophage (CD68-positive cell) infiltration. Meanwhile, ATG down- regulated the mRNA levels of pro-inflammatory chemokine monocyte chemoattractant protein-1 (MCP- 1) and cytokines, including tumor necrosis factor- α (TNF- α), interleukin-1 β (IL-1 β), and interferon- γ (IFN- γ ), in the obstructed kidneys. This was associated with decreased activation of nuclear factor κB (NF- κB). ATG attenuated UUO-induced oxidative stress by increasing the activity of renal manganese superoxide dismutase (SOD2), leading to reduced levels of lipid peroxidation. Furthermore, ATG inhibited the epithelial-mesenchymal transition (EMT) of renal tubules by reducing the abundance of transforming growth factor- β1 (TGF- β1) and its type I receptor, suppressing Smad2/3 phosphorylation and nuclear translocation, and up-regulating Smad7 expression. Notably, the efficacy of ATG in renal protection was comparable or even superior to losartan. Conclusion: ATG could protect the kidney from UUO-induced injury and fibrogenesis by suppressing inflammation, oxidative stress, and tubular EMT, thus supporting the potential role of ATG in renal fibrosis treatment. ©2017 Elsevier GmbH. All rights reserved.

Introduction

α α Abbreviations: -SMA, -smooth muscle actin; ANOVA, analysis of variance; Chronic kidney disease (CKD) has emerged as a worldwide pub- ATG, arctigenin; BCA, bicinchoninic acid; BUN, blood urea nitrogen; CAT, catalase; lic health problem with a rapid growth in prevalence ( Couser et al., CKD, chronic kidney disease; Cr, creatinine; DAB, 3, 3 -diaminobenzidine; DMSO, dimethyl sulfoxide; ECM, extracellular matrix; EMT, epithelial-mesenchymal tran- 2011). Progressive tubulointerstitial fibrosis (TIF) is the common sition; GPx, Glutathione peroxidase; GR, glutathione reductase; H 2 O 2 , hydrogen pathological presentation of nearly all kinds of CKD leading to peroxide; HYP, hydroxyproline; IFN- γ , interferon- γ ; IL-1 β, interleukin-1 β; MCP- end-stage renal failure ( Liu, 2006 ). Although tremendous efforts κ 1, monocyte chemoattractant protein-1; MDA, malondialdehyde; NF- B, nuclear have been made to prevent or retard the progression of TIF, factor- κB; PVDF, polyvinylidene difluoride; qPCR, quantitative real-time PCR; SOD1, copper-zinc superoxide dismutase; SOD2, manganese superoxide dismutase; T βR-I, TGF- β type I receptor; T βR-II, TGF- β type II receptor; TBA, thiobarbituric acid; TECs, tubular epithelial cells; TGF- β1, transforming growth factor- β1; TIF, tubulointersti- E-mail addresses: [email protected] (R. Wang), [email protected] , tial fibrosis; TNF- α, tumor necrosis factor- α; UUO, unilateral ureteral obstruction. [email protected] (P. Li).

∗ Corresponding author. 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.phymed.2017.03.003 0944-7113/© 2017 Elsevier GmbH. All rights reserved. A. Li et al. / Phytomedicine 30 (2017) 28–41 29 specific drug therapies to delay the progression of TIF toward of renal fibrosis. Thus, antioxidant and/or anti-inflammatory agent end-stage renal failure are limited. have been used to protect the kidney from injury induced by UUO. CKD can initially manifest as inflammatory responses Burdock ( Arctium lappa L.) has been used in traditional and folk characterized by infiltration of immune cells, mainly mono- medicine in oriental countries for centuries. Various experimental cytes/macrophages and T lymphocytes, into the glomeruli and models have shown that arctigenin (ATG) is a main bioactive tubulointerstitium ( Lopez-Novoa and Nieto, 2009 ). On one hand, ingredient derived from dried fruit of A. lappa . ATG exhibits these inflammatory cells secrete various pro-inflammatory cy- anti-inflammatory ( Hyam et al., 2013 ), anti-oxidant ( Zhang et al., tokines and chemokines such as tumor necrosis factor- α (TNF- α), 2015 ), and antineoplastic ( Awale et al., 2006 ) properties in var- interleukin-1 β (IL-1 β), interferon- γ (IFN- γ ), and monocyte ious disorders. Our preliminary study demonstrated that oral chemoattractant protein-1 (MCP-1), which further contribute to administration of 95% (v/v) hydroalcoholic extract from A. lappa recruitment of circulating inflammatory cells. This creates a ma- containing a higher amount of ATG and its glycoside Arctiin versus lignant positive feedback loop of inflammation ( Diamond et al., aqueous and petroleum ether extracts, exhibited the strongest 1994 ). On the other hand, activated macrophages produce profi- inhibitory effect on ECM component synthesis in the obstructive brotic cytokines such as transforming growth factor- β1 (TGF- β1), kidneys of rats with UUO (Supplementary Fig. S1). Our previous which have been shown to induce activation of matrix-producing data also showed that ATG could reverse TGF- β1-triggered re- myofibroblasts ( Leask and Abraham, 2004 ). The stimulated my- nal tubular EMT-like phenotypic changes in vitro using human ofibroblasts cause accumulation of extracellular matrix (ECM) proximal tubular epithelial cells (HK-2 cells)—mainly due to the proteins and lead to complete destruction of renal parenchyma inhibition of TGF- β1-induced up-regulation of MCP-1 ( Li et al., and irreversible renal failure. Renal inflammation—and in partic- 2015 ). Because the regulatory role of ATG in renal fibrosis in vivo ular the accumulation of macrophages in renal interstitium—are has not yet been studied, we studied here a 14-day UUO rat model a critical element in the mechanism responsible for the initia- to assess the anti-fibrotic efficacy of ATG and further delineated tion and development of renal fibrogenesis. Therefore, inhibit- the potential molecular mechanisms by which ATG elicits its ing inflammatory responses may significantly attenuate renal effects on experimental renal fibrosis. TIF. Recent studies show that a significant portion of synthetically Materials and methods active myofibroblasts arise from renal tubular epithelial cells (TECs) via the epithelial-mesenchymal transition (EMT) in fibrotic kidney Chemicals diseases ( Liu, 2010 ). The pro-fibrogenic effect of inflammation depends, at least partially, on triggering EMT ( Wynn, 2008 ). Several Arctigenin (MW: 372.41) was provided by Nanjing Zelang Med- studies have shown that sustained stimulation of pro-inflammatory ical Technology Co. Ltd. (Nanjing, China). The purity of ATG was cytokines TNF- α or IL-1 can induce EMT in the epithelial cell lines determined to be > 8% using high performance liquid chromatog- ( Takahashi et al., 2010 ). During EMT, differentiated TECs lose raphy (HPLC; Supplementary Fig. S2). Losartan was purchased their epithelial characteristics and undergo multiple biochemical from Merck Sharp & Dohme Ltd. (Hangzhou, China). Antibodies changes, which enable them to assume a mesenchymal phenotype. against the following proteins were used: NF- κB p65 (Signalway This phenotypic conversion involves the de novo synthesis of mes- Antibody, Pearland, TX, USA); fibronectin (BD Biosciences, San enchymal cytoskeletal biomarkers such as α-smooth muscle actin Jose, CA, USA); Smad2/3, phosphorylated Smad2/3 (p-Smad2/3) ( α-SMA), fibronectin and vimentin, a down-regulation of epithelial and vimentin (Cell Signaling Technology, Beverly, MA, USA); E- biomarkers such as E-cadherin and zonula occludens-1, and the cadherin, lamin B1, I κB α, TGF- β1, and TGF- β type I (T βR-I) and acquisition of a fibroblastic morphology with a concomitant inva- type II (T βR-II) receptors (Santa Cruz Biotechnology, Santa Cruz, sive phenotype. TGF- β1 has been identified as the main inducer CA , USA); α-SMA , SOD1, manganese superoxide dismutase (SOD2) of EMT and consequent interstitial ECM production in kidney and and Smad7 (Epitomics, Burlingame, CA, USA); β-actin (Bioworld other organ systems ( Acloque et al., 2009; Pan et al., 2015 ). The Technology, Minneapolis, MN, USA); collagen type I and CD68 intracellular Smad pathway is important for TGF- β1 to initiate (Boster biological technology, Wuhan, China). Serum creatinine, tubular EMT and fibrotic responses ( Lan and Chung, 2011 ). To blood urea nitrogen (BUN), urinary protein, hydroxyproline (HYP), prevent progression of TIF, the inhibition of tubular EMT mediated malondialdehyde (MDA), and SOD assay kits were obtained from by TGF- β1/Smad signaling pathway may be helpful. Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Glu- Unilateral ureteral obstruction (UUO) is a well-established tathione peroxidase (GPx), glutathione reductase (GR), CAT, and experimental model used to elucidate pathological mechanisms of bicinchoninic acid (BCA) protein assay kits were obtained from chronic obstructive nephropathy ( Chevalier et al., 2009 ). Oxidative Beyotime Institute of Biotechnology (Jiangsu, China). All other stress plays a central role in the progression of renal damage in chemicals were obtained from Sigma Chemical Co. (St. Louis, MO, obstructed nephropathy ( Kawada et al., 1999 ). Oxidative stress USA) unless otherwise specified. occurs when the production of reactive oxygen species (ROS) exceeds the capacity of intrinsic antioxidant defense. In UUO • − kidneys, the superoxide anion (O2 ) and its derivative hydrogen Animals peroxide (H2 O2 ) are increased, while the antioxidant enzyme catalase (CAT) and copper-zinc superoxide dismutase (SOD1) Eight-week-old male Sprague-Dawley rats (180–200 g; certifica- mRNA correspondingly decreased ( Ricardo et al., 1997 ). Increased tion No: SCXK-Yu-2012-0 0 05) were supplied by the Experimental ROS concentrations can cause tubulointerstitial injury by increas- Animal Center of Third Military Medical University (Chongqing, ing lipid peroxidation, hydrogen peroxides, leukocyte activation, China). All studies were performed in accordance with the Na- DNA breakdown, protein oxidation, and apoptosis. Furthermore, tional Institutes of Health guidelines for the Care and Use of ROS generated in TECs induce nuclear factor kappa B (NF- κB) Laboratory Animals (8th Edition, 2011). The experimental pro- activation ( Gloire et al., 2006 ) leading to transcription of some tocols were approved by the Institutional Ethics Committee of pro-inflammatory mediators such as MCP-1, TNF- α, INF- γ , and Chongqing University of Technology. All rats were housed in an IL-1 β, infiltration of monocytes/macrophages, proliferation of air-conditioned room at 21 ± 2 °C and 50 ± 5% relatively humidity fibroblasts, and ECM accumulation in the renal interstitium. Initial under a 12-h light/dark cycle and had access to standard rodent oxidative stress and inflammation contribute to the progression chow and water ad libitum throughout the study period. 30 A. Li et al. / Phytomedicine 30 (2017) 28–41

UUO model and treatment protocols Cybernetics, Bethesda, MD, USA). For each experimental group, 5 fields from each of 5 different Masson’s trichrome-stained sections A UUO kidney disease model was induced in rats by ligation per rat were analyzed at 200 × magnification. The areas of fibrotic of the left ureter as described previously ( Kinter et al., 1999 ). All lesion in the cortical interstitium were calculated in arbitrary rats were randomly divided into 6 groups (6 rats in each group): fields and expressed as a percentage of fibrotic area relative to (1) sham group where the rats underwent sham operations and the entire area. All semiquantitative analysis was performed in a were administered with vehicle; (2) sham + ATG group where blinded fashion. the rats underwent sham operations and were treated with ATG (3 mg/kg/d); (3) UUO model group where the rats were subjected Measurement of renal HYP content to UUO but were treated with vehicle; (4) UUO + ATG low-dose group where the rats underwent UUO and were treated with ATG The HYP content in renal tissues was measured using a com- (1 mg/kg/d); (5) UUO + ATG high-dose group where the rats under- mercial detection kit (Nanjing Jiancheng Bioengineering Institute, went UUO and were treated with ATG (3 mg/kg/d); (6) UUO + losar- Nanjing, China) according to the manufacturer’s instructions. The tan group where the rats underwent UUO and were administered results are expressed as mg of HYP per gram of tissue. with losartan (20 mg/kg/d, an angiotensin II receptor antagonist). The dosage of ATG and losartan was selected based on previous re- Immunohistochemical analysis port ( Awale et al., 2006 ) and our pilot study. The ATG and losartan were first dissolved in dimethyl sulfoxide (DMSO) at 4 or 12 mg/ml Kidney paraffin sections were deparaffinized in xylene and and kept at −20 °C. The stock solution of ATG was diluted in saline rehydrated. Non-enzymatic antigen retrieval was performed by to a final concentration of 0.1 or 0.3 mg/ml until immediately heating the sections to 98 °C in citrate buffer (10 mM, pH 6.0) before administration. The final volume of DMSO in saline is 2.5%, for 10 min. The sections were immersed into 3% (v/v) H2 O2 and this served as the vehicle. On day 4 after obstructive surgery, in methanol for 20 min at room temperature to quench poten- rats were administered vehicle, ATG, or losartan by gastric gavage tial endogenous peroxidase activity. Following three washes with once daily for 11 consecutive days. All rats were sacrificed under phosphate buffered saline (PBS; pH 7.4) after blocking, the sections pentobarbital sodium anesthesia on day 14 postoperatively. were subsequently incubated with primary antibodies overnight at 4 °C in a humidity chamber. The primary antibodies were Urine, blood and tissue preparation as follows: rabbit monoclonal antibody against α-SMA (1:100 dilution); rabbit polyclonal antibody against E-cadherin, vimentin, The day before euthanasia, the urine samples were collected p-Smad2/3, CD68, TGF- β1, and TGF- β1 receptors (1:100 dilution); from rats that were housed individually for 24 h in metabolic or mouse monoclonal antibody against fibronectin (1:100 dilu- cages to measure urinary protein excretion. Whole-blood samples tion). The final incubation was carried out for 30 min at 37 °C were collected from the inferior vena cava, and creatinine and with peroxidase-labeled polymer conjugated to goat anti-mouse BUN concentrations were analyzed from those serum samples. The or anti-rabbit immunoglobulin (Beijing Zhongshan Golden Bridge left kidney was harvested, de-capsulated, and separated into two Biotechnology Co., Ltd., Beijing, China). Immunostaining was visu- parts—one part of the renal tissues were fixed in 10% (v/v) buffered alized with hydrogen peroxide and 3, 3 -diaminobenzidine (DAB) formaldehyde and embedded in paraffin for histopathologic and as the chromogen. The sections were lightly counterstained with immunohistochemical examinations. The remaining tissues were hematoxylin and mounted on a coverslip. Negative controls were snap-frozen using liquid nitrogen and stored at −80 °C for either created by substituting the primary antibodies mentioned above protein or RNA analysis. with non-immune mouse or rabbit serum. Photomicrographs of the sections ( n = 5 per animal) at ×200 magnification were taken Renal function evaluation using an optical microscope (Olympus) and analyzed using Image- Pro Plus image analysis software. Staining intensity was computed The 24-h urinary protein excretion was measured using the as the integrated optical density (IOD). For semi-quantitative Bradford method (Nanjing Jiancheng, China). Serum concentrations analysis, the IOD was measured counting 3 randomly collected of creatinine and BUN were measured with commercially available fields with the same area for each section. In addition, the number colorimetric kits (Nanjing Jiancheng, China). These were expressed of interstitial CD68-positive macrophages and phosphorylated as μmol/L and mmol/L, respectively. Smad2/3-positive cells were assessed in 3 fields of each of 5 sections ( ×200 magnification). Histopathological examination RNA reverse transcription and quantitative real-time PCR (RT-qPCR) Paraffin-embedded sections of renal tissues (4 μm thickness) were deparaffinized and stained by hematoxylin and eosin (H&E) The total RNA was isolated from each renal cortical tissue using standard methods. Renal tubulointerstitial lesions char- sample using RNApure high-purity total RNA rapid extraction kit acterized by tubular dilation and epithelial desquamation with (spin-column, BioTeke, Beijing, China) according to the manufac- interstitial expansions were graded on a scale from 0 to 4 accord- turer’s instruction. The first strand complementary DNA (cDNA) ing to the extent of cortical involvement: 0, normal; 1, changes was synthesized from equal amounts of total RNA (2 μg) using affecting < 25% of the cortex; 2, changes affecting 25–50% of the M-MuLV reverse transcriptase with Oligo(dT) primers (BioTeke) in cortex; 3, changes affecting 50–75% of the cortex; 4, changes a 20 μl reaction volume. Real-time quantitative PCR analysis was affecting > 75% of the cortex. The overall mean scores corre- performed using SYBR FAST qPCR Master Mix (Kapa Biosystems, sponding to tubulointerstitial lesions in each rat were calculated Woburn, MA, USA) and primers ( Table 1 ). These were specific for based on individual values, which were determined on at least rat MCP-1, TNF- α, IL-1 β, IFN- γ , IL-10 and β-actin. A standard 10 randomly chosen non-overlapping fields at a magnification curve from consecutive 5-fold dilutions of a cDNA pool represen- of 200 ×. The degree of TIF was assessed by Masson’s trichrome tative of all samples was generated in each experiment to verify staining and collagen quantification. Digital images of renal tissues qPCR efficiency. The threshold cycle (Ct) is defined as the number were acquired using an optical microscope (Olympus) equipped of cycles required for the fluorescent signal to become detectable with image analysis software (Image-Pro Plus version 6.0; Media above background. The Ct levels are inversely proportional to the A. Li et al. / Phytomedicine 30 (2017) 28–41 31

Table 1 Primer sequences for quantitative real-time PCR ampliﬁcation. Gene Accession number Primer nucleotide sequence (5 –3 ) Product length (bp)

MCP-1 NM_031530.1 Forward: GCCAGATGCAGTTAATGCCC 118 Reverse: AGCTTCTTTGGGACACCTGC TNF- α NM_012675.3 Forward: AAATGGGCTCCCTCTCATCAGTTC 111 Reverse: TCCGCTTGGTGGTTTGCTACGAC IL-1 β NM_031512.2 Forward: CACCTCTCAAGCAGAGCACAG 79 Reverse: GGGTTCCATGGTGAAGTCAAC IFN- γ NM_138880.2 Forward: AACCAGGCCATCAGCAACAACA 214 Reverse: ACCGACTCCTTTTCCGCTTCCT IL-10 NM_012854.2 Forward: TTGAACCACCCGGCATCTAC 91 Reverse: CCAAGGAGTTGCTCCCGTTA β-actin NM_031144.3 Forward: CCCTGTGCTGCTCACCGA 186 Reverse: ACAGTGTGGGTGACCCCGTC amount of the initial template number. Fold changes in mRNA ex- Molecular docking pression were calculated according to the comparative Ct method (2 − Ct) as described in our previous report ( Li et al., 2015 ). The crystal structure of rat SOD is unavailable in the Pro- tein Data Bank (PDB), and thus homology models of rat SOD were built with the macromolecules module in Discovery Stuido

Western blotting analysis 3.0 software suite (Accelrys Software, http://accelrys.com/ ). The amino acid sequences of rat SOD were retrieved from Uniport

The separation of cytoplasmic and nuclear fractions using a NE- database ( http://www.uniprot.org/; SOD1: Q6LDS4, SOD2: P07895).

PER nuclear and cytoplasmic extraction kit (Pierce Biotechnology, Three-dimensional structures of human SOD were downloaded

Rockford, IL, USA) was performed according to the manufacturer’s from PDB (SOD1: 4B3E; SOD2: 1N0J). The template sequence was recommendations. The appropriate fractionation was conﬁrmed identiﬁed through rat SOD sequence on NCBI &PDB using BLAST β by Western blotting for -actin (a cytoplasmic protein) and lamin homology searching with default parameters in the Discovery

B1 (a nuclear protein). Tissues were lysed with a lysis buffer Studio version 3.0. The target and the selected template sequences

(150 mM NaCl, 50 mM Tris pH 8.0, 0.5% NP-40, 5 mM EDTA, 1 M were aligned with ClustalW. Next, molecular docking was used

DTT, 100 mM PMSF, 10 mg/ml aprotinin and leupeptin) and soni- to predict the modes of interactions of ATG with the rat SOD cated briefly to produce whole renal cortical homogenates. Protein protein. The three-dimensional structure of ATG was then energy concentration was quantified using the BCA assay method and minimized with a default force field, and partial charges were equal amounts of protein extract (50 μg/lane) were separated by calculated using MMFF94 method in SYBYL version 2.0 (Tripos

8–12% sodium dodecylsulphate-polyacrylamide gel electrophore- Int., USA). Molecular docking studies used the docking module in sis (SDS-PAGE). The separated proteins were electro-transferred LigandFit implemented in the Receptor-Ligand Interaction package to a polyvinylidene diﬂuoride (PVDF) membrane (Millipore, of the Discovery Studio version 3.0 software suite. The docking

Bedford, MA, USA). These membranes were blocked for 1 h at parameters were the default values. room temperature with 5% nonfat milk followed by incubation with primary antibodies at 4 °C overnight. After being washed Statistical analysis four times in TBS-T, each membrane was incubated for 1 h at room temperature with a 1:30 0 0 dilution of a horseradish per- Data are expressed as mean ± standard deviation (SD). Compar- oxidase (HRP)-conjugated secondary anti-rabbit or anti-mouse isons of the ranked data between different groups were performed immunoglobulin (Bioworld Technology). Immunoreactive bands using nonparametric Kruskal-Wallis one-way analysis of variance were visualized with an enhanced chemiluminescence (ECL) kit (ANOVA); when the differences were significant by this test (Millipore). Images were captured using a Gel Doc/Chemi Doc ( P < 0.05), the Mann-Whitney U test was subsequently performed Imaging System and the intensity of each band was quantified by to identify significant differences by pairwise comparison of each Quantity One software (Bio-Rad, Hercules, CA, USA). group. For other data, their statistical significance was evaluated by one-way ANOVA followed by Tukey post hoc test. All analysis Measurement of antioxidant enzyme and lipid peroxidation was completed with SPSS version 16.0 Statistical Software (IBM, Chicago, IL, USA), and P < 0.05 was considered to be statistically Renal cortical tissues were washed with saline to remove red significant. blood and clots, and were then weighed and homogenized in cold Tris-HCl (5 mM containing 2 mM EDTA, pH 7.4). After the Results homogenate was centrifuged at 1500 g for 20 min at 4 °C, aliquots of the supernatants were analyzed for lipid peroxidation and ATG attenuated progressive fibrosis in the obstructed kidneys induced antioxidant enzymes. Kidney lipid peroxidation was determined by UUO by measuring the level of MDA according to the manufacturer’s protocol. The level of MDA was expressed as μM/g protein. Antiox- To investigate whether ATG can suppress renal TIF in vivo , we idant enzymes were assayed by measuring the activities of GPx, examined interstitial fibrotic lesions and collagen accumulation in GR, CAT, and SOD. GPx and GR activities were expressed as μM UUO kidneys. As shown in Fig. 1 A, compared with sham-operated NADPH oxidized/min/g protein. The CAT activity was expressed as controls, the UUO caused obvious tubulointerstitial injury con- μ M H2 O2 consumed/min/mg protein. The hydroxylamine method sisting of tubular dilatation and epithelial atrophy, interstitial was used to determine SOD activity, which was expressed as inflammatory cell infiltration, and a marked dilation of interstitial units (U)/mg protein. One unit of SOD activity was defined as spaces. This interstitial expansion was accompanied by a noticeable the amount of enzyme that inhibited the oxidation reaction of increase in collagen deposition as shown by a collagen-specific hydroxylamine by 50%. Masson’s trichrome staining ( Fig. 1 B). However, treatment with 32 A. Li et al. / Phytomedicine 30 (2017) 28–41

Fig. 1. ATG suppressed renal interstitial ﬁbrosis in the kidney tissues of UUO rats. (A) Representative photomicrographs of hematoxylin and eosin-stained kidney sections of rats from different groups: (a) sham, (b) sham + ATG (ATG 3 mg/kg/d treatment only), (c) UUO model (UUO operated), (d) UUO + ATG (UUO operated plus ATG 1 mg/kg/d), (e) UUO + ATG (UUO operated plus ATG 3 mg/kg/d), and (f) UUO + losartan (UUO operated plus 20 mg/kg/d losartan). (B) Representative photomicrographs of kidney sections by Masson’s trichrome staining in each group of rats. Scale bar: 50 μm. (C) Tubulointerstitial damage and (D) the degree of tubulointerstitial collagen deposition were assessed semiquantitatively as described in the Materials and Methods section. (E) Total kidney collagen content was measured using the HYP assay. Each bar represents mean ± SD ∗ for 6 rats. #P < 0.05 versus sham group; P < 0.05 versus UUO model group.

ATG and losartan substantially reduced the interstitial volume ex- these findings. Fig. 2 A shows few CD68-positive macrophages in pansion and collagen deposition. Semi-quantitative morphometric the sham-operated kidneys; 14 days after UUO, a greater num- analysis revealed that ATG reduced the scores of tubulointerstitial ber of macrophages infiltrated into the tubular interstitium of injury in the obstructed kidneys at day 14 after UUO ( Fig. 1 C). obstructed kidneys than sham-operated counterparts. Treatment The administration of 1 or 3 mg/kg/d ATG led to renal interstitial with ATG or losartan prevented renal interstitial infiltration of fibrotic space that was only 47% or 14% of the levels in the UUO macrophages in the obstructed kidneys. Quantification of im- rats ( Fig. 1 D). Quantitative determination of renal HYP content munostaining is shown in Fig. 2 B indicating that the number of also confirmed a reduced collagen deposition in the cortical in- CD68-positive macrophages in the tubulointerstitium was signifi- terstitium after ATG or losartan treatment ( Fig. 1 E). There was no cantly elevated 14 days after UUO, and the increase was alleviated significant difference between ATG and losartan groups. Together, by ATG treatment in a dose-dependent manner. these findings suggest that ATG ameliorates UUO-induced renal We next examined whether ATG had an effect on the proin- fibrosis and ECM production. flammatory mediator gene expression in macrophages. The Neither ATG nor UUO treatment significantly influenced the RT-qPCR results showed that the mRNA levels of several pro- serum levels of creatinine, BUN, and urinary protein excretion inflammatory mediators such as MCP-1 ( Fig. 2 C), TNF- α ( Fig. 2 D), ( Table 2 ). This may be attributed to the traits of the UUO model IL-1 β ( Fig. 2 E) and INF- γ ( Fig. 2 F) were significantly up-regulated in which the contralateral kidney is mostly normal and sufficient in the kidneys after UUO, and their increases were significantly for maintaining renal function. attenuated by ATG or losartan treatment. Interestingly, we found that UUO kidneys displayed marked induction of IL-10 mRNA ATG relieved inflammatory responses in the obstructed kidneys versus sham controls ( Fig. 2 G), while ATG or losartan treatment did not alter UUO-induced IL-10 mRNA expression. The recruitment of macrophages to the tubulointerstitium is In resting cells, NF- κB is sequestered in the cytoplasm in a dominant event in the initiation of TIF. Immunohistochemical an inactive state through its interaction with inhibitory protein staining of macrophages used anti-CD68 antibody and confirmed I κB ( Sanz et al., 2010 ). Upon fibrogenic stimuli, the I κB pro- A. Li et al. / Phytomedicine 30 (2017) 28–41 33

Table 2 Biochemical parameters of rats among different groups.

Group Serum creatinine Serum blood urea nitrogen Urinary protein (n = 6) ( μmol/L) (mmol/L) (mg/day)

Sham 61 .74 ± 6.84 6.38 ± 0.95 15 .21 ± 5.74 ATG (3 mg/kg/d) 64 .73 ± 7.59 6.10 ± 1.07 11 .14 ± 4.30 UUO model 65 .94 ± 7.99 6.75 ± 1.29 17 .11 ± 9.39 UUO + ATG (1 mg/kg/d) 62 .87 ± 8.95 7.12 ± 1.02 12 .87 ± 7.32 UUO + ATG (3 mg/kg/d) 63 .10 ± 5.93 7.30 ± 0.64 8.75 ± 4.86 UUO + losartan (20 mg/kg/d) 69 .64 ± 4.94 6.74 ± 1.28 9.31 ± 4.50 teins are phosphorylated, ubiquitinated, and degraded, thereby ATG regulated TGF- β1/Smad signaling pathway in the obstructed releasing NF- κB to enter the nucleus and trigger transcription of kidneys inflammation-related cytokine and chemokine genes. To clarify whether ATG or losartan inhibited activation of NF- κB, we mea- TGF- β1-induced EMT in a variety of cells is mediated by sured relative levels of I κB α in the cytoplasm and NF- κB p65 in activation of serine/threonine kinase receptors and canonical Smad the nucleus by Western blotting, respectively ( Fig. 2 H and I). In the signaling pathway ( Lan and Chung, 2011 ). We first evaluated kidneys of sham-operated rats, there were very high levels of cy- the effect of ATG on renal TGF- β1, T βR-I and T βR-II abundance toplasmic I κB α and basal levels of nuclear NF- κB p65 proteins. after obstructive injury. Immunohistochemical staining for TGF- β1 In response to UUO, the NF- κB was significantly activated, as localized the expression of TGF- β1 on kidney sections. In kidneys reflected by a simultaneous decrease in cytoplasmic I κB α and at day 14 after UUO, clear tubular expression of TGF- β1 was an increase in nuclear NF- κB p65. Conversely, after treatment observed ( Fig. 4 A and D) in addition to its expression in interstitial with ATG or losartan, the I κB α was retained in the cytoplasm, areas. In contrast, ATG or losartan treatment significantly inhibited and almost no NF- κB p65 signal was found in the nuclear fraction TGF- β1 protein expression in the treatment groups 14 days after suggesting that ATG or losartan notably enhanced stability of I κB UUO. We also examined the expression of T βR-I and T βR-II in and blocked nuclear translocation of NF- κB. the obstructed kidneys. Renal expression of T βR-I ( Fig. 4 B and These results collectively describe how ATG inhibits UUO- D) and T βR-II ( Fig. 4 C and D) was markedly up-regulated in induced renal inflammatory responses—the compound inhibits the obstructed kidneys at day 14 versus sham-operated kidneys. UUO-induced macrophage infiltration, NF- κB activation and nu- Interestingly, the expression of T βR-I and T βR-II protein was pri- clear translocation. This prevents pro-inflammatory mediator marily induced in renal tubules. After ATG or losartan treatment, expression in macrophages. the up-regulated tubular expression of T βR-I in diseased kidneys was reduced, whereas protein expression of T βR-II was almost ATG inhibited tubular EMT in the obstructed kidneys unchanged. Consistent with the immunohistochemistry, Western blotting analysis indicated that TGF- β1 and T βR-I expression Whether renal inflammation progresses to TIF depends in part was significantly up-regulated in UUO rats compared with sham on whether TECs undergo a process of EMT. To determine the control group ( Fig. 4 E and F). These changes were significantly effects of ATG on UUO-mediated changes related to EMT after attenuated in ATG or losartan-treated UUO rats. inflammatory responses, the expression of EMT markers in the We next investigated the effect of ATG on activation status obstructed kidneys was assessed using immunohistochemical and and expression levels of Smad proteins in the obstructed kidneys. Western blotting analysis. In kidney sections from sham-operated After obstructive injury, Smad2/3 signaling was highly activated rats, epithelial marker E-cadherin displayed moderate or strong as revealed by higher levels of phosphorylated and total Smad2/3 complete membranous staining in all tubular epithelial cells protein ( Fig. 5 A and B), but significantly lower levels of inhibitory ( Fig. 3 A). On the other hand, the mesenchymal marker α-SMA was Smad7 protein ( Fig. 5 A and C) than in sham-operated kidneys. found exclusively in the vascular smooth muscle cells ( Fig. 3 B). By immunohistochemical staining, we examined the localization Two other markers of mesenchymal cells—fibronectin ( Fig. 3 C) and of p-Smad2/3 in kidneys of rats with UUO on day 14. As shown vimentin ( Fig. 3 D)—were almost not detected by immunostaining. in Fig. 5 D, p-Smad2/3 was predominantly localized in the nuclei Compared with sham-operated controls, expression of E-cadherin of severely congested tubular epithelial cells but also intersti- protein decreased in renal cortical areas of the UUO rats ( Fig. 3 E), tial cells. The quantification of immunostaining also showed an especially around the renal tubules; conversely, expression levels increasing number of p-Smad2/3-positive nuclei in renal cortical of α-SMA ( Fig. 3 F), fibronectin ( Fig. 3 G) and vimentin ( Fig. 3 H) area after UUO ( Fig. 5 E) indicating Smad2/3 nuclear localization proteins were significantly elevated in the UUO rats suggesting in the fibrotic kidneys. Treatment with ATG or losartan notably that the TECs were undergoing EMT. The ATG administration attenuated UUO-induced phosphorylation and nuclear location of largely blocked induction of α-SMA, fibronectin and vimentin ex- Smad2/3, which was associated with a significant up-regulation pression in kidney tissues of UUO rats while markedly inhibiting a of renal Smad7. These results indicate that ATG might suppress UUO-induced decrease in E-cadherin expression. Losartan had the UUO-induced TIF by regulation the TGF- β1/Smad pathway. same effects. Consistent with the immunohistochemistry studies, Western blotting analysis also showed that the UUO model re- ATG reduced MDA content, increased activity and expression of SOD2, sulted in a marked decrease in the E-cadherin/ α-SMA ratio ( Fig. 3 I but had no effect on activities of GPx, GR, CAT, and SOD1 in the and G) concomitant with an increase in fibronectin and vimentin obstructed kidneys protein levels ( Fig. 3 I and K). In contrast, the expression of α-SMA, fibronectin and vimentin was down-regulated with ATG or losartan Oxidative stress was evaluated by determination of MDA levels treatment in the obstructed kidneys of rats with UUO for 14 days, and activities of the antioxidant enzymes in the obstructed kid- but the expression of E-cadherin increased in response to ATG or neys. Fig. 6 A shows a significant increase in MDA production in the losartan treatment. These data suggest that ATG exerts its anti- UUO-injured kidneys accompanied by a corresponding decrease fibrotic effect through inhibition of EMT-like phenotypic changes in GPx ( Fig. 6 B), GR ( Fig. 6 C), and CAT ( Fig. 6 D) activities. ATG or and preservation of the normal tubular epithelial phenotype. losartan administration significantly decreased the MDA content 34 A. Li et al. / Phytomedicine 30 (2017) 28–41

Fig. 2. ATG regulated inflammatory signaling pathway in the kidney tissues of UUO rats. (A) Representative photomicrographs illustrating infiltration of CD68-positive macrophages (black color directed by arrows) in the kidney tissues of rats in different groups: (a) sham (sham operated), (b) sham + ATG (ATG 3 mg/kg/d treatment only), (c) UUO model (UUO operated), (d) UUO + ATG (UUO operated plus ATG 1 mg/kg/d), (e) UUO + ATG (UUO operated plus ATG 3 mg/kg/d), and (f) UUO + losartan (UUO operated plus 20 mg/kg/d losartan). Scale bar: 50 μm. (B) Quantification of the number of CD68-positive macrophages per high-power field in different groups. (C–G) RT-qPCR analysis of gene expression was performed for inflammatory mediators MCP-1, TNF- α, IL-1 β, INF- γ , and IL-10. (H) Cytoplasmic and nuclear fractions of kidney cortex tissues were collected and analyzed by Western blotting to detect I κB α and NF- κB p65, respectively. The β-actin and lamin B1 protein levels served as internal controls, respectively, for cytoplasmic extracts (CE) and nuclear extracts (NE). (I) The protein levels of I κB α in cytoplasmic extracts and NF- κB p65 in nuclear extracts were converted to arbitrary densitometric units normalized by the value of the corresponding loading control and expressed relative to the protein level in sham-operated group (defined as 1-fold). ∗ Each bar represents mean ± SD of three independent experiments. n = 6 rats per group. #P < 0.05 versus sham group; P < 0.05 versus UUO model group.

in kidney tissues, while UUO-induced reduction of renal GPx, SOD1 activity was markedly decreased. ATG or losartan treatment GR, and CAT activities was not ameliorated by ATG or losartan further enhanced the activity of SOD2 when compared with the treatment. UUO model group. Interestingly, treatment with ATG alone in rats The activity of antioxidant enzyme SOD2 was signiﬁcantly without UUO operation increased the activity of SOD2. However, elevated in the obstructed kidneys of UUO rats ( Fig. 6 E), while ATG had no effect on the activity of SOD1 in sham or UUO- A. Li et al. / Phytomedicine 30 (2017) 28–41 35

Fig. 3. ATG inhibited the EMT process in the kidney tissues of UUO rats. (A–D) The location and expression of E-cadherin, α-SMA, fibronectin and vimentin were determined by immunohistochemical staining in the kidney sections of rats from different groups: (a) sham (sham operated), (b) sham + ATG (ATG 3 mg/kg/d treatment only), (c) UUO model (UUO operated), (d) UUO + ATG (UUO operated plus ATG 1 mg/kg/d), (e) UUO + ATG (UUO operated plus ATG 3 mg/kg/d), and (f) UUO + losartan (UUO operated plus 20 mg/kg/d losartan). Scale bar: 50 μm. (E–H) Semi-quantitative immunohistochemical analysis of the EMT-related protein expression in different groups. (I) Kidney tissue lysates were studied by Western blotting analysis with specific antibodies against E-cadherin, α-SMA, fibronectin, vimentin or β-actin. β-Actin was used as an internal control. (G, K) The protein levels of E-cadherin, α-SMA, fibronectin and vimentin were expressed as arbitrary densitometric units and normalized by the value of β-actin and expressed relative to the protein levels and ratio of E-cadherin/ α-SMA protein levels in the sham-operated group (defined as 1-fold). Each bar represents mean ± SD of ∗ three independent experiments. n = 6 rats per group. #P < 0.05 versus sham group; P < 0.05 versus UUO model group. 36 A. Li et al. / Phytomedicine 30 (2017) 28–41

Fig. 4. Effects of ATG on TGF- β1, T βR-I and T βR-II protein expression in the kidney tissues of UUO rats. (A–C) Representative photomicrographs of immunohistochemical staining for TGF- β1, T βR-I, and T βR-II in the kidney sections of rats from different groups: (a) sham (sham operated), (b) sham + ATG (ATG 3 mg/kg/d treatment only), (c) UUO model (UUO operated), (d) UUO + ATG (UUO operated plus ATG 1 mg/kg/d), (e) UUO + ATG (UUO operated plus ATG 3 mg/kg/d), and (f) UUO + losartan (UUO operated plus 20 mg/kg/d losartan). Scale bar: 50 μm. (D) Semi-quantitative immunohistochemical analysis of the TGF- β1-related protein expression in different groups. (E) Kidney tissue lysates were subject to Western blotting analysis with antibodies against TGF- β1, T βR-I, T βR-II, or β-actin. β-Actin was used as an internal control. (F) The protein levels of TGF- β1, T βR-I, and T βR-II were converted to arbitrary densitometric units and normalized by the value of β-actin and finally expressed according to the levels ∗ in sham-operated group (defined as 1-fold). Each bar represents mean ± SD of three independent experiments. n = 6 rats per group. #P < 0.05 versus sham group; P < 0.05 versus UUO model group. operated kidneys. Consistent with the results of SOD1 and SOD2 ATG interacted with SOD in docking study activities, Western blotting studies showed that ATG treatment further increased UUO-induced up-regulation of SOD2 expression The ATG molecule was docked into the three-dimensional at day 14 after surgery ( Fig. 6 F and G), but did not significantly structures of SOD to envisage the potential interaction ( Fig. 7 A change the renal expression of the SOD1. and B). It is interesting to note that ATG offered a relatively higher docking score for the binding pocket of rat SOD2 (dock_score: 81.688) than SOD1 (dock_score: 80.383) implying that more A. Li et al. / Phytomedicine 30 (2017) 28–41 37

Fig. 5. Effects of ATG on Smad family members in the kidney tissues of UUO rats. (A) Representative Western blotting photographs showing protein levels of p-Smad2/3, Smad2/3, Smad7, and β-actin in the kidney tissue lysates. The protein levels of Smad2/3 and β-actin were used as internal controls, respectively. (B, C) The protein levels of p-Smad2/3 and Smad7 were converted to arbitrary densitometric units, normalized by the value of the corresponding loading control and expressed relative to the phosphorylation ratio or to the protein levels in sham-operated rats (deﬁned as 1-fold). (D) The kidney sections were subjected to immunohistochemical staining with antibodies against p-Smad2/3. Representative photomicrographs of p-Smad2/3 immunoreactivity are shown in different groups: (a) sham (sham operated), (b) sham + ATG (ATG 3 mg/kg/d treatment only), (c) UUO model (UUO operated), (d) UUO + ATG (UUO operated plus ATG 1 mg/kg/d), (e) UUO + ATG (UUO operated plus ATG 3 mg/kg/d), and (f) UUO + losartan (UUO operated plus 20 mg/kg/d losartan). (E) Graphical depiction of number of p-Smad2/3-positive cells per high-power ﬁeld in the kidney sections of ∗ rats from different groups. Scale bar: 50 μm. Each bar represents mean ± SD of three independent experiments. n = 6 rats per group. #P < 0.05 versus sham group; P < 0.05 versus UUO model group.

binding contacts and/or lower energy and steric penalties are cal- ing dialysis or kidney transplantation ( Liu, 2006 ). The activation culated. The results of the molecular dockings showed functional of inflammatory cascade is an early and characteristic feature of molecular interactions between ATG and SOD. As shown in Fig. 7 C, CKD. It is widely accepted that sustained inflammation initiates ATG was predicted to interact with SOD1 at the following sites: the progression of renal injury in CKD, which eventually leads to Asp10; Gly 9, and 55; Ala 144; Arg 142; Cys 56, and 145; Thr the destruction and collapse of the renal parenchyma replaced 53, and 57; and Gln 54. The molecule was also found to bind by fibrotic scar tissues. An intrinsic regulatory loop connecting with the following SOD2 amino acid residues: His 54, and 187; inflammation and TIF is that damaged TECs recruit inflammatory Trp 185; Tyr 190; and Leu 191 ( Fig. 7 D). Compared with SOD1, cells into the renal interstitial compartments and, in turn, infil- the SOD2/ATG complex had more hydrophobic interactions with trated inflammatory cells—especially macrophages—also produce a aromatic residues. Furthermore, a hydrogen bonding interaction myriad of pro-fibrotic and pro-inflammatory cytokines, and then was predicted between the hydroxyl group of the lignans’ 4- act on renal TECs and resident fibroblasts to facilitate TIF in a hydroxy-3-methoxyphenyl moiety and the residue at His 187 of paracrine fashion ( Leask and Abraham, 2004 ). This hypothesis is SOD2 ( Fig. 7 E). There was no hydrogen bond formed between ATG confirmed by the available literature and our own studies as re- and SOD1, which might account for the potential affinity of ATG flected by macrophage infiltration found in the obstructed kidneys for SOD2 but not SOD1. These data suggested that ATG could bind after UUO using CD68 staining; nuclear translocation of NF- κB and to SOD2 and confirmed that SOD2 could be a molecular target for up-regulation of the recruitment signal MCP-1 was also seen in the ATG enhancement of antioxidant capacity. UUO kidneys. Infiltrating macrophages in UUO typically correlate with the degree of renal fibrosis, and disruption of macrophage recruitment reduces fibrosis in several disease models. We found Discussion that ATG limited the extent of renal inflammation in UUO, which possibly resulted from inhibition of NF- κB activity, reduced ex- Renal TIF has emerged as a common feature of progressive CKD because it eventually progresses to end-stage renal failure requir- 38 A. Li et al. / Phytomedicine 30 (2017) 28–41

Fig. 6. ATG attenuated UUO-induced oxidative stress. (A) Kidney tissues were homogenized and subjected to TBA-reactive substances assay for the measurement of the lipid peroxidation product MDA. Proteins were isolated from kidney tissues and subjected to enzymatic assays for the determination of (B) GPx, (C) GR, (D) CAT and (E) SOD activities. (F) Proteins were extracted from kidney tissues and subjected to Western blotting analysis using primary antibodies speciﬁc for SOD1 and SOD2, respectively. Blots were stripped and re-probed with antibodies against β-actin to correct for differences in protein loading. (G) Fold-changes in the relative levels of SOD1 and SOD2 protein are shown after normalizing with β-actin. The relative protein level in the sham group was deﬁned as 1.0. Each bar represents mean ± SD of three independent experiments. ∗ n = 6 rats per group. #P < 0.05 versus sham group; P < 0.05 versus UUO model group.

pression of MCP-1, and the resulting amelioration of macrophage IFN- γ polarize macrophages towards the M1 phenotype which infiltration in the tubular interstitium. produces large amounts of TNF- α, IL-1 β, IL-6, IL-12, IL-23, and Recent studies have demonstrated the diversity and plasticity MCP-1, increased levels of reactive nitrogen, and reactive oxygen of macrophage phenotypes and functionality ( Duffield, 2011 ). intermediates. However, exposure of macrophages to the T helper The macrophage M1/M2 polarization shift may cause different 2 cytokines IL-4/IL-13 produces a M2 phenotype that is charac- functions and even opposite effects in the progression of kidney terized by efficient phagocytic activity, production of ornithine diseases. Macrophages are heterogeneous population of immune and polyamines through the arginase pathway, and up-regulated cells that may acquire distinct functional phenotypes in response expression of the mannose and galactose receptors, IL-10 and to different stimuli. Lipopolysaccharide and T helper 1 cytokine TGF- β1. Increasing evidence suggests that classically activated M1 A. Li et al. / Phytomedicine 30 (2017) 28–41 39

Fig. 7. Modeled interactions of ATG with SOD1 or SOD2. (A, B) The docking binding pocket of ATG with SOD1 or SOD2 after homology modeling. (C, D) Amino acid residues of SOD1 or SOD2 involved in ATG binding. (E) Key hydrogen bonding interaction (purple line) between ATG and His 187 of SOD2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) macrophages are implicated in initiating and sustaining inflamma- the macrophage subpopulations could inter-switch their phenotype tion while M2 macrophages are associated with anti-inflammatory over the course of the response to UUO. One to four days after and homeostatic functions linked to inflammation resolution and UUO, macrophage activation was skewed to the M1 type but later tissue remodeling ( Wang et al., 2007 ). This study showed that on, M2 macrophages were the major class of macrophages, which obstructed kidneys expressed high levels of M1 cytokines TNF- α, deteriorated renal fibrosis ( Fujiu et al., 2011 ). Our study found that IL-1 β and IFN- γ . The level of M2 macrophages cytokines IL-10 in addition to its inhibitory properties against pro-inflammatory and TGF- β1 were also up-regulated simultaneously 14 days after M1 cytokine expression, ATG dramatically inhibited expression UUO injury indicating that an adoptive M1 to M2 macrophage of the M2 cytokine TGF- β1. Nevertheless, another M2 marker shift takes place at later stages of UUO. IL-10, whose mRNA was not differentially expressed in obstructed Earlier studies suggested that adoptive transfer of macrophages kidneys after ATG treatment. This discrepancy warrants additional ameliorated renal fibrosis ( Nishida et al., 2005 ). However, more investigation but may reflect the differential effect of ATG during recent studies showed that M2 macrophage depletion specifically a sequential course of macrophage phenotype transition. Thus, our inhibited EMT and subsequent interstitial fibrosis ( Pan et al., 2015 ). observations provided a mechanistic explanation for ATG shifting These results were consistent with recent reports suggesting that 40 A. Li et al. / Phytomedicine 30 (2017) 28–41 the inflammatory milieu toward resolution by modulation of a A recent report indicates that ROS play an important role macrophage subset polarization. in UUO-induced inflammatory responses through activation of Over the past decade, many studies have identified tubular NF- κB pathway ( Qin et al., 2016 ). ROS mediate TGF- β1-induced EMT as a critical event in the pathogenesis and progression of EMT through ERK-directed activation of Smad pathway in TECs renal TIF. Therefore, targeting EMT has been proposed as an ( Rhyu et al., 2005 ). These results support a role for ROS as potent effective therapeutic strategy for preventing renal fibrosis. Our signaling molecules. Inhibition of ROS generation by activation of experiments demonstrate that E-cadherin—a key component of scavenging enzymes may prevent renal inflammation and EMT in adhesion junctions and critical in the maintenance of epithelial obstructive kidneys. The exacerbated UUO-induced TIF is associ- integrity—was highly re-expressed by treatment with ATG, but ated with overproduced ROS in the obstructed kidneys accompa- expression of the mesenchymal markers α-SMA, fibronectin and nied by decreased antioxidant enzyme activities including SOD and vimentin was strongly down-regulated. These data suggest that CAT ( Ricardo et al., 1997 ). Here, we demonstrate for the first time ATG may inhibit myofibroblast and ECM formation, at least in part, that ATG exerted obvious renoprotective effects through decreasing through inhibition of TEC transdifferentiation. levels of MDA—a known product of lipid peroxidation in UUO rats. We know that TGF- β1, the major isoform of the TGF- β family, This suggests that the anti-fibrogenic effect of ATG is associated can be secreted by all types of renal cells and infiltrated in- with reduced oxidative stress. Notably, we found that treatment flammatory cells. Consistent with the concept, we noted TGF- β1 with ATG significantly increased SOD2 activity in sham- or UUO- in TECs and the renal interstitium. Interestingly, ATG treatment operated rat kidneys. This effect was specific to SOD2 because ATG significantly down-regulated protein expression of TGF- β1 in the had no obvious effects on other antioxidant enzymes including injured kidneys, and this is accompanied by attenuation of inter- GPx, GR, CAT and SOD1. Next, we further investigated whether the stitial macrophage infiltration. Reduced TGF- β1 expression by ATG increase in SOD2 activity resulted from an up-regulation of SOD2 may be the indirect consequence of reduced fibrogenic cellular expression. Western blotting showed that the renal protein level components including myofibroblasts that originate from TECs by of SOD2 remained unchanged in sham-manipulated rats treated EMT. Also, this decrease in the level of TGF- β1 by ATG may be with ATG alone despite the increased activity of SOD2 under the attributed to fewer inflammatory cells such as macrophages in same conditions. This suggests that increased activity of SOD2 by the interstitium of the UUO kidney; TGF- β1 itself is a potential ATG does not merely result from up-regulation of its expression. chemoattractant for these cell types. It was shown recently that Hence, we next employed molecular docking studies to predict M2 macrophages, not myofibroblast, are the main source for the interactions between ATG and SOD2. Our current docking this profibrogenic cytokine ( Pan et al., 2015 ). Indeed, we found results revealed that ATG established a hydrogen bond with that ATG decreased the number of infiltrating and M2-polarized important active residue His 187 of SOD2. In addition to hydro- macrophages, which were identified by immuohistochemistry and gen bonding, several hydrophobic interactions between ATG and quantitative analysis of RT-PCR. aromatic residues of SOD2 were observed, e.g. , His 54, and 187; The canonical signaling pathway induced by TGF- β1 involves Trp 185; Tyr 190; and Leu 191. The hydrogen and hydrophobic activation of the cellular substrates Smad2 and Smad3 by ser- interactions of ATG to the active site residues might anchor the ine/threonine phosphorylation through the T βR-I kinase ( Lan and molecule on the active sites of SOD2 leading to their conforma- Chung, 2011 ). The activated Smad2/3 then binds to co-activator tional changes and subsequent increase in enzyme activity. Thus, Smad4 to form a complex, which subsequently translocates into our data provide valuable insight into the molecular basis of ATG. the nucleus where it regulates the transcription of EMT and Further experiments are required to validate SOD2 as a direct fibrosis-related target genes. On the other hand, activated T βR-I molecular target for ATG. also induces the expression of Smad7 via a Smad3-dependent In this study, we employed losartan, an angiotensin II re- mechanism. This inhibitory Smad7 serves as an antagonist of ceptor blocker, as a positive control. It is widely accepted that TGF- β1 signaling by binding to the activated receptor complex angiotensin II plays a central role in the initiation and progression to prevent access and phosphorylation of Smad2/3. In addition, of renal inflammation, TGF- β production and EMT ( Esteban et al., Smad7 functions as an adaptor protein to recruit E3 ligases Smurf2 2004 ). Here, we found that losartan markedly mitigated the UUO- and Arkidia to the T βR-I to promote its degradation through the induced kidney fibrotic changes including interstitial infiltration of ubiquitin proteasome pathway ( Kavsak et al., 20 0 0 ). At the same macrophages, tubular atrophy, tubulointerstitial space expansion time, ubiquitin-degradation of Smad7 occurs simultaneously. Here, and ECM protein accumulation (fibronectin, vimentin and type we provide evidence that UUO-induced renal fibrosis was associ- I collagen). It is noteworthy that the effect of ATG (3 mg/kg/d) ated with a significant phosphorylation and nuclear translocation and losartan (20 mg/kg/d) on major determinants of renal TIF are of Smad2/3 in the obstructed kidneys, while Smad7 was reduced. similar, albeit at different doses. In addition, no mortality and In contrast, the ATG treatment effectively reversed these changes. obvious treatment-related clinical signs of toxicity were observed More importantly, other investigators have previously demon- in animals treated with ATG (3 mg/kg/d) alone during the as- strated that reduced levels of inhibitory Smad7 protein caused by sessment period (14 days) suggesting that ATG may be relatively Smurf2-dependent degradation may be the underlying mechanism safe when used to treat renal fibrosis. required for promotion of TGF- β/Smad-mediated renal fibrosis. In contrast, overexpression of renal Smad7 could directly induce the Conclusion expression of I κB to inhibit NF- κB activation and NF- κB-driven inflammatory responses ( Wang et al., 2005 ). These results sug- For the first time, we evaluate the anti-fibrotic effect of ATG gest that blockade of TGF- β/Smad signaling by enhancing the using the UUO model. We found that ATG could attenuate tubular expression level of Smad7, but not influence of Smad2/3, reduces injuries and accumulation of ECM proteins in the kidneys of UUO activation of NF- κB and restores the imbalance between TGF- rats. We also noted that ATG could suppress the infiltration of β/Smad signaling. This may be a key mechanism by which ATG macrophages via inactivation of NF- κB pathway in the progression substantially inhibited the UUO-induced inflammatory responses, of renal TIF while inhibiting pro-inflammatory cytokine (TNF- α, Smad2/3 activation, and the resulting EMT. Gene knockdown IL-1 β, and INF- γ ) and chemokine (MCP-1) release. Meanwhile, studies of Smad7 mRNA are therefore warranted to provide further ATG could block the TGF- β1-induced tubular EMT process, which evidence of the renoprotective properties of ATG by induction of was likely mediated via regulation of the TGF- β1/Smad pathway. Smad7 in the development of renal TIF. Our findings also revealed that ATG might protect against renal in- A. Li et al. / Phytomedicine 30 (2017) 28–41 41 juries by increasing the enzyme activity of SOD2 and subsequently Kavsak, P. , Rasmussen, R.K. , Causing, C.G. , Bonni, S. , Zhu, H. , Thomsen, G.H. , reducing renal oxidative stress. Molecular docking investigations Wrana, J.L. , 20 0 0. Smad7 binds to Smurf2 to form an E3 ubiquitin ligase that targets the TGF β; receptor for degradation. Mol. Cell 6, 1365–1375 . indicated that the SOD2 could be a target molecule for ATG effects. Kawada, N. , Moriyama, T. , Ando, A. , Fukunaga, M. , Miyata, T. , Kurokawa, K. , Imai, E. , These findings might be used for further investigations to explore Hori, M. , 1999. Increased oxidative stress in mouse kidneys with unilateral the therapeutic potential of ATG in the attenuation of progressive ureteral obstruction. Kidney Int. 56, 1004–1013 .

Kinter, M., Wolstenholme, J.T., Thornhill, B.A., Newton, E.A., McCormick, M.L., Cheva- renal TIF. lier, R.L. , 1999. Unilateral ureteral obstruction impairs renal antioxidant enzyme activation during sodium depletion. Kidney Int. 55, 1327–1334 . Conflict of interest Lan, H.Y. , Chung, A.C. , 2011. Transforming growth factor- β; and Smads. Contrib. Nephrol. 170, 75–82 . Leask, A. , Abraham, D.J. , 2004. TGF- β; signaling and the fibrotic response. FASEB J. The authors declare no competing financial interest. 18, 816–827 . Li, A. , Wang, J. , Zhu, D.P. , Zhang, X.X. , Pan, R.L. , Wang, R. , 2015. Arctigenin suppresses β Acknowledgments transforming growth factor- ;1-induced expression of monocyte chemoattractant protein-1 and the subsequent epithelial-mesenchymal transition through reactive oxygen species-dependent ERK/NF- κ;B signaling pathway in renal This work was supported by the National Natural Science tubular epithelial cells. Free Radic. Res. 49, 1095–1113 . Foundation of China (No. 81001675 ), Macao Science and Technol- Liu, Y. , 2006. Renal fibrosis: new insights into the pathogenesis and therapeutics. Kidney Int. 69, 213–217 . ogy Development Fund (No. 007/2014/AMJ), Research Committee Liu, Y. , 2010. New insights into epithelial-mesenchymal transition in kidney fibrosis. of the University of Macau (No. MYRG2014-0 0 089-ICMS-QRCM ), J. Am. Soc. Nephrol. 21, 212–222 . Chongqing Natural Science Foundation (No. cstc2014jcyjA10030 ). Lopez-Novoa, J.M. , Nieto, M.A. , 2009. Inflammation and EMT: an alliance towards organ fibrosis and cancer progression. EMBO Mol. Med. 1, 303–314 . Nishida, M. , Okumura, Y. , Fujimoto, S. , Shiraishi, I. , Itoi, T. , Hamaoka, K. , 2005. Adop- Supplementary materials tive transfer of macrophages ameliorates renal fibrosis in mice. Biochem. Bio- phys. Res. Commun. 332, 11–16 .

Pan, B.X., Liu, G.H., Jiang, Z.P., Zheng, D.W., 2015. Regulation of renal fibrosis by Supplementary material associated with this article can be macrophage polarization. Cell. Physiol. Biochem. 35, 1062–1069 . found, in the online version, at doi:10.1016/j.phymed.2017.03.003 . Qin, T. , Du, R.H. , Huang, F.J. , Yin, S.S. , Yang, J. , Qin, S.Y. , Cao, W.S. , 2016. Sinomenine activation of Nrf2 signaling prevents hyperactive inflammation and kidney in- References jury in a mouse model of obstructive nephropathy. Free Radic. Biol. Med. 92, 90–99 . Rhyu, D.Y. , Yang, Y. , Ha, H. , Lee, G.T. , Song, J.S. , Uh, S.T. , Lee, H.B. , 2005. Role of re- Acloque, H. , Adams, M.S. , Fishwick, K. , Bronner-Fraser, M. , Nieto, M.A. , 2009. Epithe- active oxygen species in TGF- β;1-induced mitogen-activated protein kinase ac- lial-mesenchymal transitions: the importance of changing cell state in develop- tivation and epithelial-mesenchymal transition in renal tubular epithelial cells. ment and disease. J. Clin. Invest. 119, 1438–1449 . J. Am. Soc. Nephrol. 16, 667–675 . Awale, S. , Lu, J. , Kalauni, S.K. , Kurashima, Y. , Tezuka, Y. , Kadota, S. , Esumi, H. , 2006. Ricardo, S.D. , Ding, G. , Eufemio, M. , Diamond, J.R. , 1997. Antioxidant expression in Identification of arctigenin as an antitumor agent having the ability to eliminate experimental hydronephrosis: role of mechanical stretch and growth factors. J. the tolerance of cancer cells to nutrient starvation. Cancer Res. 66, 1751–1757 . Am. Soc. Nephrol. 272, F789–F798 . Chevalier, R.L. , Forbes, M.S. , Thornhill, B.A. ,2009. Ureteral obstruction as a model Sanz, A.B. , Sanchez-Nino, M.D. , Ramos, A.M. , Moreno, J.A. , Santamaria, B. , of renal interstitial fibrosis and obstructive nephropathy. Kidney Int. 75, Ruiz-Ortega, M. , Egido, J. , Ortiz, A. , 2010. NF- κ;B in renal inflammation. J. Am. 1145–1152 . Soc. Nephrol. 21, 1254–1262 . Couser, W.G. , Remuzzi, G. , Mendis, S. , Tonelli, M. , 2011. The contribution of chronic Takahashi, E. , Nagano, O. , Ishimoto, T. , Yae, T. , Suzuki, Y. , Shinoda, T. , Nakamura, S. , kidney disease to the global burden of major noncommunicable diseases. Kid- Niwa, S. , Ikeda, S. , Koga, H. ,2010. Tumor necrosis factor- α; regulates transform- ney Int. 80, 1258–1270 . ing growth factor- β;-dependent epithelial-mesenchymal transition by promot- Diamond, J.R. , Kees-Folts, D. , Ding, G. , Frye, J.E. , Restrepo, N.C. , 1994. Macrophages, ing hyaluronan-CD44-moesin interaction. J. Biol. Chem. 285, 4060–4073 . monocyte chemoattractant peptide-1, and TGF-beta 1 in experimental hy- Wang, W. , Huang, X.R. , Li, A.G. , Liu, F. , Li, J.H. , Truong, L.D. , Wang, X.J. , Lan, H.Y. , dronephrosis. Am. J. Physiol. 266, F926–F933 . 2005. Signaling mechanism of TGF- β;1 in prevention of renal inflammation: Duffield, J.S. , 2011. Macrophages in kidney repair and regeneration. J. Am. Soc. role of Smad7. J. Am. Soc. Nephrol. 16, 1371–1383 . Nephrol. 22, 199–201 . Wang, Y. , Wang, Y.P. , Zheng, G. , Lee, V.W. , Ouyang, L. , Chang, D.H. , Mahajan, D. , Esteban, V. , Lorenzo, O. , Ruperez, M. , Suzuki, Y. , Mezzano, S. , Blanco, J. , Kretzler, M. , Coombs, J. , Wang, Y.M. , Alexander, S.I. , Harris, D.C. , 2007. Ex vivo programmed Sugaya, T. , Egido, J. , Ruiz-Ortega, M. , 2004. Angiotensin II, via AT1 and AT2 re- macrophages ameliorate experimental chronic inflammatory renal disease. Kid- ceptors and NF- κ;B pathway, regulates the inflammatory response in unilateral ney Int. 72, 290–299 . ureteral obstruction. J. Am. Soc. Nephrol. 15, 1514–1529 . Wynn, T. , 2008. Cellular and molecular mechanisms of fibrosis. J. Pathol. 214, Fujiu, K. , Manabe, I. , Nagai, R. , 2011. Renal collecting duct epithelial cells regulate in- 199–210 . flammation in tubulointerstitial damage in mice. J. Clin. Invest. 121, 3425–3441 . Zhang, W.Z. , Jiang, Z.K. , He, B.X. , Liu, X.B. , 2015. Arctigenin protects against Gloire, G. , Legrand-Poels, S. , Piette, J. , 2006. NF- κ;B activation by reactive oxygen lipopolysaccharide-induced pulmonary oxidative stress and inflammation in a species: fifteen years later. Biochem. Pharmacol. 72, 1493–1505 . mouse model via suppression of MAPK, HO-1, and iNOS signaling. Inflamma- Hyam, S.R. , Lee, I.A. , Gu, W. , Kim, K.A. , Jeong, J.J. , Jang, S.E. , Han, M.J. , Kim, D.H. , tion 38, 1406–1414 . 2013. Arctigenin ameliorates inflammation in vitro and in vivo by inhibiting the PI3K/AKT pathway and polarizing M1 macrophages to M2-like macrophages. Eur. J. Pharmacol. 708, 21–29 .